Science vs. Politics

I just read an excellent essay by Michael Crichton about the subversion of science by politics.

Posted by dustin on December 15, 2004 with category tags of

5 comments
Wow, that's an awesome speech, relevant in too many ways to count. Thanks very much for the link!
   comment by tentons on December 16, 2004

Great article, I read it and forwarded the link to my friend Kevin Yager who some of you might know. All I said was "read, reply" and he did, extensively. Read his thoughts here

A highly interesting article, needless to say. I largely agree with Michael
Crichton (hereafter MC), even if he does overstate many things. It is
always dangerous for someone to write something so aggressive and
attacking, since inevitable his own arguments should be placed under the
same scrutiny. For instance, in his analysis he says things like:

"The result: it took the consensus fifty years to acknowledge what any
schoolchild sees."

"Now. You tell me you can predict the world of 2100. Tell me it's even worth
thinking about. Our models just carry the present into the future. They're
bound to be wrong. Everybody who gives a moment's thought knows it."

Here he commits both scientific errors that he hates: he makes wild
conclusions without actually using facts or scientific analysis, and he
appeals to consensus to prove that he's right.

He claims forward-projecting climate models are wrong, but then uses a
classic "look at these past errors/trends... clearly those also apply today
and in the future" logic (which is really the same flimsy debating).

So basically, he's no better than all the rest of them. Like all of us,
once he believes in something, he'll use whatever argument seems handy to
make his point.I think many of his arguments are dumb and should be
discarded. Some are not so dumb.

The central one, of saying that in some undefined way SETI and Global
Warming and all this are connected, is clearly dumb. I know he's using it
for laughs, but ultimately even insinuating a conspiracy when trying to be
so serious is counterproductive.


Let me pick at some of the things he said:

Drake Equation: I do agree that this equation lacks substance. I've seen it
many times and always wondered why it was held to be so important when
it's really just the simplest statement of statistics. Yet MC goes a bit
overboard in saying it's totally useless. If it is used properly, it can
set bounds on how likely alien life is, as we come to have better estimates
of the various parameters. And some scientists are trying to pin down
those values. For instance, the search for extra-solar planets has made at
least one of the
parameters go from being "completely unknown" to "we have some limit on
the number of planets out there... there are some!"

Here, as in many other places, MC claims that there is no point in coming
up with an analysis, when in fact what he means is "people should honestly
report error bars." The Drake equation is perfectly fine; it's just that
the values are so unknown that the final estimate becomes: “one billion +/-
one billion billion”; i.e.: the prediction is as good as useless, but the
framework isn't necessarily.

Similarly elsewhere, if climate models were always honest with error bars:
"we predict a temperature increase of 10 C +/- 21 C," then everyone would
know just how useless it is to establish policy based on the prediction. I
think even lay people could the learn pretty quickly what an error interval
is, in the same way that they sorta understand other simple statistics.

MC's statements that:
"SETI is unquestionably a religion."
are neither true nor helpful. I agree it's flimsy science, but the main
mistake they are making, in my opinion, is under-reporting their error
intervals. Many scientists, like me, believe that aliens exist (but have
not contacted Earth of course), but would never bet on it. We feel that the
evidence indicates that they probably exist, but we acknowledge that the
errors are much too large for it to be anything more than speculation at
this point. In religion, the belief is absolute (or it's supposed to
be). In science (as in most SETI-believers), the belief is not
absolute. We all acknowledge that it may or may not be true.

Here again is the difference with SETI and creationism. If creationism (or
rather "intelligent design") is just another scientific theory, then that's
fine: we'll analyze it and accept it if it explains observations or makes
predictions. It's the fact that creationists want absolute acceptance of
their beliefs that makes it religious instead of scientific. I don't think
quite the same claims can be made about SETI. Except for certain alien-
worshipping cultists, no SETI person is going to try and force you to
believe aliens exist. They just want to know if aliens exist or not: that's
just scientific curiosity.

Again with the nuclear winter equation he says that it's useless because
the values are not well known. Again I say that he should instead argue
using error bars instead of just saying that the whole attempt is
misguided. Militaries and governments need models that can provide ranges
of possibilities. If the range is "maybe we'll be fine and maybe we'll all
die" then they will be more worried than if the range is "maybe we won't even
notice and maybe we'll basically be fine after 2 years" (even though, in
both cases, the range of possibilities is huge) So I disagree with MC that
models with inherent ambiguities are not worth exploring. Knowing the range
of possibilities is helpful.

Consensus: I totally agree with MC that science is not based on consensus:
it's based on verifiable facts and evaluation of predictions.

Still, I think there is a subtle difference between "Science" and
"science." Let's say that "Science" is what MC is talking about, and it's
what I really love: the pursuit of truth (with the embedded unjustifiable
philosophical belief that our predictive theories are actual
representations of reality). "science" on the other hand, is what most
people care about: it's what makes their TVs work and their computers fast
and tells them whether a
bridge is going to collapse or not.

For the average joe on the street, who doesn't understand "Science" or
"science", he wants some metric of whether some prediction is valid or not.
So he appeals to consensus: how many scientists believe in such-and-such a
prediction? It's a flawed system, but at least it gives the joe some
indication of how likely some bit of "science" is to be true. (This
analysis is based on the idea that the joe cannot verify the science
himself with reading or experimentation; he's just an average joe after
all.) Even though MC points out how often the scientific consensus is
wrong, it is still a logical
strategy for joe to use the consensus as a measure of the likely truth of a
scientific statement, since he cannot verify the statements himself, and
since most of the time a "sceptic" is in fact a "crackpot." I think it's
reasonable for policy makers to use the "consensus of scientists" argument.

It is, however, clearly wrong for a scientist to try and convince another
scientist using the consensus argument. In that instance it has no
weight. The scientist should just present the experimental observations
and the theory, and all other scientists should be able to clearly see if
the theory is right or wrong. However, to policy makers and joes, it makes
a bit more sense to be swayed by consensus.

It's similar to the way that most people accept what the doctor says after
getting a second opinion: the doctors may be wrong, but statistically if
lots of doctors agree on a given thing, then they are probably right.
Putting your faith in someone who breaks with medical consensus is not
intelligent. If the marginalized person is actually right, then eventually
their theory will be vindicated. That in practice it takes hundreds of
years for the consensus to change doesn't change the argument on a per-case
basis.

So really what MC should argue for is for SCIENTISTS to analyze claims
without accepting "consensus" as an argument. So that then the claim that
"scientists agree that..." will carry the weight it deserves. In real
science, this is usually what happens. Scientists very much enjoy
challenging the establishment, and conversely enjoy proving that the
established ideas are correct. In practice, it usually works.

And even in the scientific establishment, it's been shown that a certain
amount of orthodoxy and resistance makes science robust against crackpots,
fraudulent data, and mistakes. If some fact is established by 100
experiments, and suddenly refuted by 1, we shouldn't change our worldview
until that 1 refutation has been reproduced and expanded upon.

MC is right that scientists should not abuse the privileged position of respect
that they (sorta) have in society by manipulating science to enforce their
personal views. They should let the science (and common sense) speak for
itself.

I've been saying it for awhile: even if the data on pollution do not
unambiguously show that we have altered the climate, we should STILL reduce
pollution and emissions, and so on. The mere fact that our emissions are on
the right order-of-magnitude to have some effect on the environment is
cause for concern, and should affect our policies.

So we shouldn't have to manipulate and abuse the science for people to get
their act together. Scientists should clearly say: "we don't know for sure
that pollution will destroy the earth..." And then remind people that even
a 10% chance of bad stuff happening is worth worrying about. For example,
if you had a 10% chance of dying after eating a given food, would you eat
it? (Even if it tasted really good?) I think scientists should make it
clear that what's going on is Russian Roulette: we have no clue what we are
doing, and there is a fairly substantial % chance that what we are doing is
going to seriously screw us over.

MC is being rather idealistic about science, considering that he's talking
so much about public policy. I agree with him that science should never
overstate its knowledge. But he should really put this into perspective
for lay people. Most lay people think that scientists are arrogant in their
beliefs; they don't understand that the true ideal of a scientist is a
person who is able to change their worldview (theories) once new evidence
arrives, and that scientists are the FIRST people to point out that
ultimately we can't be sure of anything.

In court cases (at least on "Law and Order" I mean), they always question
the doctors or scientists with things like "but you can't be sure,
right?" To which the scientist dutifully replies "no I can't be sure of
that." But it's unfair to compare the scientific version of certainty with
the lay version of certainty. A normal person, under oath, will swear that
they are "sure" that they saw the person wearing a blue shirt, whereas a
scientist will say they are "unsure" of whether atoms or gravity really exist.

What is my point? I'm again going on about how when linking science to
policy, error bars have to be emphasized. Moreover, the public must come
to understand to some extent what scientists mean by certainty and
confidence. I think MC is ignoring this, when in my view it is the critical
missing piece. Scientists wouldn't have to exaggerate to the public if the
public understood what we meant when we used statistics and exact language.
I think most scientist want to do the right thing, but end up doing the
unethical thing (exaggerating, lying) only because they want the public to
understand what they are trying to say. You could even argue that the
scientific phrase "I'm 90% sure" translates in normal english to be "I'm
totally sure" and phrases like "pollution may be having an impact on the
world, but it's not clear" translates into "you should be worried about
pollution impacting the world…"

Until we take steps to bridge the gap between scientific talk, policy-maker
talk, and lay-person talk, this problem will remain. Too often the onus is
placed entirely on scientists: “dumb down your arguments so that lay people
can understand (but under no circumstance are you to lie or misreport
anything!).” In practice I think policy-makers and the public will have to
meet us part-way, and learn a bit about our language.

I love MC's use of horses/horseshit to point out how dumb
forward-predictions can be, even if it is, ultimately, a pointless
argument. Again, he's basing his prediction of the irrelevancy of an
argument based on an extrapolation of analogous irrelevancy in previous
time periods. He's using the exact same faulty logic to demonstrate how
faulty their logic is!

And for the record, what climatologists are doing with their models is to
predict what climate will be like "if things stay as they are." The whole
point is to model what will happen if we do NOTHING, so that we can know if
we should do something. Of course, some models build in a time-varying
response of society, and so on. Here again I think that MC is misguided in
being defeatist and saying the models are useless. Once again error bars
would help, but as I said, even knowing the range of possible outcomes for
the year 2100 is helpful.

That having been said, I know full well that climate modelling is bullshit,
and that basically it's not the sane argument against pollution. A sane
argument is: I hate the smell of car exhaust. Similarly for smoking: you
don't have to prove with 99% confidence that second-hand smoke causes lung
cancer to know that you get a headache after being in a smoky bar for too
long. People want scientific evidence to back up their common sense
sometimes; but for some issues (smoking, pollution, etc.) there are just so
many good reasons to change policy, we shouldn't need another scientific
one to push us over the edge. We should enact change immediately anyway.

I agree with MC's love of Science, and his desire to keep it from just being a
political tool. Overall I think many of his points are good, even if he has no
suggestions whatsoever for how to improve things.

   comment by KingCasey (#194) on December 18, 2004

hmmm Dustin, I used the "lessthan"split"greaterthan" syntax after the little intro to that post but it seems not to have put in the cut like I thought it would. Am I doing something wrong or does this only work for original posts?
   comment by KingCasey (#194) on December 18, 2004

The "split" tag only has effect on the original posts. There's no other page to split it to for comments, so it just gets all displayed here. That's fine though. Thanks for posting his insight. It's a little nit-picky, but the conclusion works well. He should track down MC's email address and send it.
   comment by dustin (#1) on December 18, 2004

This week's nature talks about
http://realclimate.org/

It's a blog being run by 9 climate researchers who want the world (mostly the media) to hear up-to-date opinions and debates from experts. Their intention is to educate the people and hold the media to a higher standard of scientific accuracy than has so far been the case.

In particular, they respond to Michael Crichton’s many doubts and complaints:
http://www.realclimate.org/index.php?p=74
   comment by ky on December 23, 2004

   

VorgTag Cloud

Written by dustin
Latest Photo
Quote of Now:
Friends
Popular Posts
Computer Games

Hey You! Subscribe to dustin's RSS feed.
Or get wider opinion in the Vorg All Author feed.

 
 

Members login here.
© Vorg Group.